Context files (AGENTS.md) and skills were only fetched from the
workspace on the first turn or when the agent changed. On subsequent
turns, stale content from persisted messages was used. This meant that
if AGENTS.md or skills were modified on the workspace between turns, the
agent wouldn't see the changes until the user created a new chat.
## Changes
- Extract `fetchWorkspaceContext` from `persistInstructionFiles` to
allow fetching workspace context without persisting
- On subsequent turns, re-fetch fresh context from the workspace instead
of reading stale persisted content; falls back to persisted messages if
the workspace dial fails
- Update `ReloadMessages` callback to re-derive instruction and skills
from reloaded database messages after compaction, instead of using
captured closure variables
- Add `formatSystemInstructionsFromParts` helper to build system
instructions directly from agent parts without requiring separate
OS/directory params
- Add tests for the new helper
<details><summary>Implementation Notes</summary>
### Root cause
In `runChat`, the `else if hasContextFiles` branch (subsequent turns)
called `instructionFromContextFiles(messages)` which read stale content
from persisted DB messages. The `ReloadMessages` callback
(post-compaction) also used captured `instruction`/`skills` closure
variables from the start of the turn, never re-deriving them.
### Approach
1. **Extract `fetchWorkspaceContext`** — Pure refactor of the fetch-only
part of `persistInstructionFiles` (agent connection, context config
retrieval, content sanitization, metadata stamping). Returns parts +
skills without persisting.
2. **Subsequent turns**: Instead of reading from persisted messages,
launch a `g2` goroutine that calls `fetchWorkspaceContext` to get fresh
context from the workspace. Falls back gracefully to persisted messages
if the workspace is unreachable.
3. **ReloadMessages**: Re-derive `instruction` from
`instructionFromContextFiles(reloadedMsgs)` and `skills` from
`skillsFromParts(reloadedMsgs)` using the freshly loaded messages, with
fallback to captured values if the reloaded messages don't contain
context (e.g. compacted away).
</details>
> 🤖 Generated by Coder Agents
## Summary
Adds `--ai-gateway-allow-byok` deployment option to control whether
users can use Bring Your Own Key (BYOK) mode with AI Gateway.
When disabled (`--ai-gateway-allow-byok=false`), BYOK requests are
rejected with a 403 and a message directing the admin to enable the
flag. Centralized key authentication works regardless of this setting.
Defaults to `true` (BYOK allowed).
---------
Co-authored-by: Danny Kopping <danny@coder.com>
## Problem
`resolveDeploymentSystemPrompt` was called inside `InTx` closures in
both `CreateChat` (`coderd/x/chatd/chatd.go`) and
`createChildSubagentChatWithOptions` (`coderd/x/chatd/subagent.go`).
That method uses `p.db` (the root store) internally to call
`GetChatSystemPromptConfig`, which requires a second DB pool checkout
while the transaction already holds one connection.
Under concurrent chat creation load (e.g., the chat scaletest at 4800
chats), this causes pool starvation: every in-flight create holds one
connection and blocks waiting for another, leading to `idle in
transaction` pileups and cascading timeouts across the entire coderd DB
pool — including unrelated background work like prebuild metrics and the
chat acquire loop.
## Fix
Move the `resolveDeploymentSystemPrompt` call before `p.db.InTx(...)` in
both call sites. The system prompt config is a read-only
deployment-level setting that does not need transactional consistency
with the chat insert, so fetching it before the transaction is both safe
and preferable (it also shortens transaction lifetime).
## Backporting
The `CreateChat` instance of this bug is also present on `release/2.32`
(`coderd/x/chatd/chatd.go` line 907). The `subagent.go` instance is not
— the child-subagent-chat creation path with its own `InTx` was added
after the branch cut.
This should be backported, but because this is only in the chat creation
path, and that's not typically hit with a great deal of concurrency in
the real world, I don't think an urgent patch for 2.32 is necessary.
## Lint gap
The existing `InTx` ruleguard rule in `scripts/rules.go` catches direct
outer-store usage (`p.db.GetFoo()`) and passing the outer store as a
function argument inside `InTx` closures, but it explicitly cannot catch
indirect access through receiver methods like
`p.resolveDeploymentSystemPrompt()` — the rule documents this blind spot
at line 273. Catching this class of bug would require interprocedural
analysis (following the callee's body to see if it touches `p.db`),
which is beyond what ruleguard's AST pattern matching can express. We're
considering a lightweight custom `go/analysis` analyzer (similar to
`paralleltestctx`) that does 1-level same-package callee inspection to
detect this pattern. In the meantime, this PR adds guidance to
`AGENTS.md` so AI reviewers can flag the pattern during code review.
This change reuses the authenticated subject's existing organization
membership information during chat creation instead of issuing an
`OrganizationMembers` query.
The current query is still correct, so this is not required for
correctness. However, `workspaceapps` already answers the same question
more cheaply from the request's RBAC subject. This extracts that logic
into `rbac.Subject.HasOrganizationMembership` and reuses it in both
places, removing an extra database lookup from chat creation without
changing the authorization behavior.
I'm currently debugging a Coder agents scaletest regression where a run
on April 2, 2026 with 4800 concurrent chat creations passed, while the
same run on April 15, 2026 does not. We could stagger chat creation to
reduce the burst, but I'd rather understand why this bottleneck appeared
in the first place so we can keep making small hot-path improvements
like this one instead of only smoothing over the symptom.
_Disclaimer: produced by Claude Opus 4.6_
Adds a `coder_build_info` metric which allows operators to see which
versions of Coder are currently running.
---------
Signed-off-by: Danny Kopping <danny@coder.com>
Addresses review findings from #23827 that were added post-merge:
- Persisted attachments now store `organizationId`; mismatched orgs
pruned on restore
- Workspace selection reconciliation: stale IDs from previous orgs
dropped via derived `effectiveWorkspaceId`
- Org picker uses `permittedOrganizations()` for RBAC-aware filtering
- Org picker hidden when user belongs to only one org
- Ref-sync `useEffect` replaced with `useEffectEvent`
- `CreateWorkspace()` and `ListTemplates()` take `organizationID` and
`db` as required function parameters instead of optional struct fields —
compiler enforces them, removes scattered nil guards
- Cross-org template check in `CreateWorkspace` is now unconditional
- `ListTemplates` org-scoping filter now has test coverage
- `setupChatInfra` comment fixed; test helpers use params structs
instead of positional UUIDs
- Enterprise test documents that org admin only sees own chats (handler
hardcodes `OwnerID` — future work needs sidebar UI before lifting that
restriction)
> 🤖
_Disclaimer: produced mostly by Claude Opus 4.6 following detailed
planning._
## Summary
- Support multiple instances of the same AI Bridge provider type via
indexed env vars (`CODER_AIBRIDGE_PROVIDER_<N>_<KEY>`), following the
`CODER_EXTERNAL_AUTH_<N>_<KEY>` pattern
- Existing single-provider env vars (`CODER_AIBRIDGE_OPENAI_KEY`, etc.)
continue to work unchanged
- Setting both a legacy env var and an indexed provider with the same
name errors at startup to prevent silent misconfiguration
- Mark legacy provider fields (`OpenAI`, `Anthropic`, `Bedrock`) as
deprecated in `AIBridgeConfig` in favor of `Providers`
## Example
```sh
CODER_AIBRIDGE_PROVIDER_0_TYPE=anthropic
CODER_AIBRIDGE_PROVIDER_0_NAME=anthropic-corp
CODER_AIBRIDGE_PROVIDER_0_KEY=sk-ant-corp-xxx
CODER_AIBRIDGE_PROVIDER_0_BASE_URL=https://llm-proxy.internal.example.com/anthropic
CODER_AIBRIDGE_PROVIDER_1_TYPE=anthropic
CODER_AIBRIDGE_PROVIDER_1_NAME=anthropic-direct
CODER_AIBRIDGE_PROVIDER_1_KEY=sk-ant-direct-yyy
```
Each instance is routed by name:
- /api/v2/aibridge/**anthropic-corp**/v1/messages
- /api/v2/aibridge/**anthropic-direct**/v1/messages
Closes
[AIGOV-157](https://linear.app/codercom/issue/AIGOV-157/spike-to-understand-if-there-is-a-simple-way-to-handle-multi-api-key)
---------
Signed-off-by: Danny Kopping <danny@coder.com>
We noticed during higher active workspace counts that the agent
connection metric, generated via a query to the database, would report a
relatively high amount of agents as disconnected. Somewhere between 5
and 20%. However, other metrics such as # of websocket connections would
suggest that all agent connections are healthy.
Looking at the `Agents` function in prometheus metrics, plus the query
execution time (not accounting for actual database RT time) revealed
that this reporting of agents as disconnected was almost certainly false
positives due to clock drift in the way we're generating the metric
values. At 10k metrics, with a p50 of 2ms and p99 of 5ms, the entire
`agents` function could take upwards of 50s to execute. Because we were
doing a query/database RT to query th apps for each agent individually,
and grabbing a `time.Now` value on each iteration of that loop, it's
likely the portion of agents that were reported as disconnected were
those that had last heartbeat the furthest in the past.
The fix here is to set a consistent `now` before fetching agent data to
avoid clock drift inflating the inactive timeout comparison, and replace
the per-agent app query N+1 with a single batched lookup to prevent loop
execution time from pushing agents over the disconnected threshold.
Signed-off-by: Callum Styan <callumstyan@gmail.com>
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
Previously we reserved some env vars that may collide with AI gateway.
These were incomplete and take away flexibility from the user, which
we're prioritizing in the first iteration of the feature.
## Summary
Add `coderd_agents_first_connection_seconds` histogram metric that
records the
duration from workspace agent creation to first connection. This fills
an
observability gap — provisioner job timings and startup script metrics
exist,
but the agent connection phase (which can take several minutes) was not
exposed
to Prometheus.
Closes https://github.com/coder/coder/issues/21282
## Changes
- **`coderd/prometheusmetrics/prometheusmetrics.go`** — Define and
register a
`HistogramVec` in the existing `Agents()` polling loop. Observe
`first_connected_at - created_at` exactly once per agent via a
deduplication
map, pruned each tick to prevent unbounded memory growth.
- **`coderd/prometheusmetrics/prometheusmetrics_test.go`** — Update
`TestAgents`
to set `first_connected_at` on the test agent and assert the histogram
is
collected with correct labels, sample count, and sample sum.
- **`docs/admin/integrations/prometheus.md`**,
**`scripts/metricsdocgen/generated_metrics`** —
Auto-generated documentation updates from `make gen`.
## Metric details
| Property | Value |
|---|---|
| Name | `coderd_agents_first_connection_seconds` |
| Type | histogram |
| Labels | `template_name`, `agent_name`, `username`, `workspace_name` |
| Buckets | 1s, 10s, 30s, 1m, 2m, 5m, 10m, 30m, 1h |
## Example PromQL
```promql
# P95 agent connection time by template
histogram_quantile(0.95,
sum(rate(coderd_agents_first_connection_seconds_bucket[1h])) by (le, template_name)
)
```
<details>
<summary>Implementation notes</summary>
### Design decisions
- **Histogram over gauge**: Enables `histogram_quantile()` for
percentile queries.
- **Observe in `Agents()` polling loop**: All required data is already
fetched by
`GetWorkspaceAgentsForMetrics()` — no new DB queries.
- **Dedup via `map[uuid.UUID]struct{}`**: Prevents re-observing the same
agent
across polling ticks. Pruned each cycle to bound memory.
- **Buckets**: Aligned with
`coderd_provisionerd_workspace_build_timings_seconds`
range (1s–1h).
### Overhead at scale (100k active workspaces)
The deduplication map (`observedFirstConnection`) and per-tick pruning
map
(`currentAgentIDs`) are both `map[[16]byte]struct{}`. At 100k agents:
- **Memory**: ~2.25 MB persistent + ~2.25 MB transient per tick = **~4.5
MB peak**.
- **CPU**: ~25 ms of map operations per tick (one tick per minute) =
**<0.05% of one core**.
Both are negligible relative to the existing cost of the `Agents()` loop
(the DB
query, per-agent `GetWorkspaceAppsByAgentID` calls, and coordinator node
lookups
dominate).
</details>
> 🤖 Generated by Coder Agents
Three SQL queries (`GetUserGroupSpendLimit`,
`ResolveUserChatSpendLimit`, `GetUserChatSpendInPeriod`) aggregated chat
spend limits and usage globally across all organizations. A restrictive
group limit in org A would bleed into org B.
## Changes
- Add `organization_id` parameter to all three SQL queries in
`coderd/database/queries/chats.sql`
- When nil UUID is passed, queries fall back to global behavior
(backward compat for HTTP dashboard endpoints)
- When real org ID is passed, limits and spend are scoped to that
organization
- Thread `organizationID` through `ResolveUsageLimitStatus` →
`checkUsageLimit` → all chatd call sites
- Update dbauthz wrappers for new param structs
- HTTP endpoints (`chatCostSummary`, `getMyChatUsageLimitStatus`) pass
`uuid.Nil` with TODO for future org-scoped UI
- Add `TestResolveUsageLimitStatus_OrgScoped` with 5 test cases covering
org isolation, nil-UUID fallback, spend scoping, and user override
priority
Closescoder/internal#1466
> 🤖
> This PR was authored by Mux on behalf of Mike.
Chats sharing one workspace (e.g. sibling subagents) all wrote to
`/home/coder/PLAN.md`, causing plan file collisions. This change derives
a unique plan path per chat from the workspace home directory and chat
ID.
## Changes
* `write_file`, `edit_files`, and `propose_plan` reject any `plan.md`
variant (case-insensitive) at the workspace home root, with a clear
error pointing to the chat-specific path.
* Root chats receive a `<plan-file-path>` block inlined in the main
system prompt with the concrete path.
* Prompt and tool descriptions no longer hardcode `/home/coder/PLAN.md`.
* Plan path handling is POSIX-only (forward-slash), relying on the
contract that workspace agent paths are normalized before reaching
chatd.
* Updated `ProposePlanTool.stories.tsx` to use per-chat path examples.
* Full test coverage for plan path detection, legacy-path rejection in
all three tools, inline prompt rendering, and fallback behavior.
* Removes experiment `web-push`.
* Falls back to NoopWebpusher in case of error
* Checks browser capability in FE
* Adds note to agents getting-started docs regarding webpush without TLS
> 🤖
## Problem
`TestWatchChats/DiffStatusChangeIncludesDiffStatus` flakes with:
```
failed to read JSON message: failed to get reader: context deadline exceeded
```
Tracker: https://github.com/coder/internal/issues/1467
## Root Cause
The test published a pubsub event **once**, immediately after
`client.Dial()` returned. However, `Dial` completes as soon as the HTTP
upgrade finishes — before the server-side `watchChats` handler has
called `SubscribeWithErr` on the pubsub channel. When the publish races
ahead of the subscription, the message is silently dropped and the
`wsjson.Read` loop blocks until the context deadline.
## Fix
Publish the event on a short ticker (`testutil.IntervalFast` = 25ms) in
a background goroutine instead of publishing once. This guarantees that
at least one publish lands **after** the subscription is active,
regardless of goroutine scheduling. The goroutine is cleanly stopped via
a `publishDone` channel once the expected event is received.
## Verification
- `go test -race -count=50 -run
TestWatchChats/DiffStatusChangeIncludesDiffStatus` → **50/50 PASS, 0
races**
- `go test -race -count=5 -run TestWatchChats` → **all 5 subtests pass
5/5 times**
## Summary
Adds a `sanitizeCredentialHint` safety check in the db-to-SDK conversion
layer to ensure credential hints are always masked before being exposed
in the API. Also adds `credential_kind` and `credential_hint` assertions
to the session threads API test.
Add secret value validation to reject null bytes and values exceeding 32KB.
The 32KB limit applies uniformly to both env var and file secrets because the
value field is shared and the destination can change after creation.
Add file path validation to also reject null bytes and paths exceeding 4096
bytes.
Wire up secret value validation into both POST and PATCH handlers.
Fixes https://github.com/coder/internal/issues/1436
* Adds organization_id to chats with backfill (workspace org → user org membership → default org)
* No support yet for ACLs (follow-up issue)
- Cross-org workspace binding rejected (both in `CreateChatRequest` and in `create_workspace` tool
- Adds `OrganizationAutocomplete` to `AgentCreateForm`
- Docs updated with `organization_id` in chats-api.md
> 🤖 Written by a Coder Agent. Reviewed by many humans and many agents.
---------
Co-authored-by: Mathias Fredriksson <mafredri@gmail.com>
The GetChats SQL query ordered by (updated_at, id) DESC with no
pin_order awareness. A pinned chat with an old updated_at could
land on page 2+ and be invisible in the sidebar's Pinned section.
Add a 4-column ORDER BY: pinned-first flag DESC, negated pin_order
DESC, updated_at DESC, id DESC. The negation trick keeps all sort
columns DESC so the cursor tuple < comparison still works. Update
the after_id cursor clause to match the expanded sort key.
Fix the false handler comment claiming PinChatByID bumps updated_at.
Closes#16332
Previously `coder provisioner jobs list` showed no indication of what a workspace
build job was doing (i.e., start, stop, or delete). This adds
`workspace_build_transition` to the provisioner job metadata, exposed in
both the REST API and CLI. Template and workspace name columns were also
added, both available via `-c`.
```
$ coder provisioner jobs list -c id,type,status,"workspace build transition"
ID TYPE STATUS WORKSPACE BUILD TRANSITION
95f35545-a59f-4900-813d-80b8c8fd7a33 template_version_import succeeded
0a903bbe-cef5-4e72-9e62-f7e7b4dfbb7a workspace_build succeeded start
```
The "By model" and "Pull requests" tables on the PR Insights page
(`/agents/settings/insights`) were side-by-side at `lg` breakpoints, and
the Pull requests table was hard-capped at 20 rows by the backend.
- Replaced `lg:grid-cols-2` with a single-column stacked layout so both
tables span the full content width.
- Removed the `LIMIT 20` from the `GetPRInsightsRecentPRs` SQL query so
all PRs in the selected time range are returned.
- Can add this back if we need it. If we do, we should add a little
subheader above this table to indicate that we're not showing all PRs
within the selected timeframe.
- Added client-side pagination to the Pull requests table using
`PaginationWidgetBase` (page size 10), matching the existing pattern in
`ChatCostSummaryView`.
- Renamed the section heading from "Recent" to "Pull requests" since it
now shows the full set for the time range.
<img width="1481" height="1817" alt="image"
src="https://github.com/user-attachments/assets/0066c42f-4d7b-4cee-b64b-6680848edc68"
/>
> 🤖 PR generated with Coder Agents
When a devcontainer subagent is terraform-managed, the provisioner sets
its directory to the host-side `workspace_folder` path at build time. At
runtime, the agent injection code determines the correct
container-internal
path from `devcontainer read-configuration` and sends it via
`CreateSubAgent`.
However, the `CreateSubAgent` handler only updated `display_apps` for
pre-existing agents, ignoring the `Directory` field. This caused
SSH/terminal
sessions to land in `~` instead of the workspace folder (e.g.
`/workspaces/foo`).
Add `UpdateWorkspaceAgentDirectoryByID` query and call it in the
terraform-managed subagent update path to also persist the directory.
Fixes PLAT-118
<details><summary>Root cause analysis</summary>
Two code paths set the subagent `Directory` field:
1. **Provisioner (build time):** `insertDevcontainerSubagent` in
`provisionerdserver.go`
stores `dc.GetWorkspaceFolder()` — the **host-side** path from the
`coder_devcontainer` Terraform resource (e.g. `/home/coder/project`).
2. **Agent injection (runtime):**
`maybeInjectSubAgentIntoContainerLocked` in
`api.go` reads the devcontainer config and gets the correct
**container-internal**
path (e.g. `/workspaces/project`), then calls `client.Create(ctx,
subAgentConfig)`.
For terraform-managed subagents (those with `req.Id != nil`),
`CreateSubAgent`
in `coderd/agentapi/subagent.go` recognized the pre-existing agent and
entered
the update path — but only called `UpdateWorkspaceAgentDisplayAppsByID`,
discarding the `Directory` field from the request. The agent kept the
stale
host-side path, which doesn't exist inside the container, causing
`expandPathToAbs` to fall back to `~`.
</details>
> [!NOTE]
> Generated by Coder Agents
Go's html/template has a built-in security filter (urlFilter) that only
allows http, https, and mailto URL schemes. Any other scheme gets
replaced with #ZgotmplZ.
The OAuth2 app's callback URL uses custom URI scheme which the filter
considers unsafe. For example the Coder JetBrains plugin exposes a
callback URI with the scheme jetbrains:// - which was effectively
changed by the template engine into #ZgotmplZ. Of course this is not an
actual callback. When users clicked the cancel button nothing happened.
The fix was simple - we now wrap the apps registered callback URI into
htmltemplate.URL. Usually this needs some validation otherwise the
linter will complain about it. The callback URI used by the Cancel logic
is actually validated by our backend when the client app
programmatically registered via the dynamic OAuth2 registration
endpoints, so we refactored the validation around that code and re-used
some of it in the Cancel handling to make sure we don't allow URIs like
`javascript` and `data`, even though in theory these URIs were already
validated.
In addition, while testing this PR with
https://github.com/coder/coder-jetbrains-toolbox/pull/209 I discovered
that we are also not compliant with
https://www.rfc-editor.org/rfc/rfc6749#section-4.1.2.1 which requires
the server to attach the local state if it was provided by the client in
the original request. Also it is optional but generally a good practice
to include `error_description` in the error responses. In fact we follow
this pattern for the other types of error responses. So this is not a
one off.
- resolves#20323
<img width="1485" height="771" alt="Cancel_page_with_invalid_uri"
src="https://github.com/user-attachments/assets/5539d234-9ce3-4dda-b421-d023fc9aa99e"
/>
<img width="486" height="746" alt="Coder Toolbox handling the Cancel
button"
src="https://github.com/user-attachments/assets/acab71a6-d29c-4fa9-80ba-3c0095bbdc8f"
/>
<!--
If you have used AI to produce some or all of this PR, please ensure you
have read our [AI Contribution
guidelines](https://coder.com/docs/about/contributing/AI_CONTRIBUTING)
before submitting.
-->
Add the five REST endpoints for managing user secrets, SDK client
methods, and handler tests.
Endpoints:
- `POST /api/v2/users/{user}/secrets`
- `GET /api/v2/users/{user}/secrets`
- `GET /api/v2/users/{user}/secrets/{name}`
- `PATCH /api/v2/users/{user}/secrets/{name}`
- `DELETE /api/v2/users/{user}/secrets/{name}`
Routes are registered under the existing `/{user}` group with
`ExtractUserParam`. The delete query was changed from `:exec` to
`:execrows` so the handler can distinguish "not found" from success
(DELETE with `:exec` silently returns nil for zero affected rows).
## Summary
Exposes `credential_kind` and `credential_hint` on AI Bridge session
threads, making credential metadata visible in the session detail API.
Each thread in the `/api/v2/aibridge/sessions/{session_id}` response now
includes:
- `credential_kind`: `centralized` or `byok`
- `credential_hint`: masked credential (e.g. `sk-a...pgAA`)
Values are taken from the thread's root interception.
## Changes
- `codersdk/aibridge.go`: Added `CredentialKind` and `CredentialHint`
fields to `AIBridgeThread`
- `coderd/database/db2sdk/db2sdk.go`: Populated from root interception
in `buildAIBridgeThread`
- `SessionTimeline.stories.tsx`: Added fields to mock thread data
The startup-timeout integration tests in `chatloop` used a 5ms real-time
budget and relied on wall-clock scheduling to fire the startup guard
timer before the first stream part arrived. On loaded CI runners the
timer sometimes lost the race, producing `attempts == 2` instead of
`attempts == 1` and flaking `TestRun_FirstPartDisarmsStartupTimeout`.
Replace the real `time.Timer` in `startupGuard` with a `quartz.Timer` so
tests can control time deterministically. Production behavior is
unchanged: `RunOptions.Clock` defaults to `quartz.NewReal()` when nil,
and the startup timeout still covers both opening the provider stream
and waiting for the first stream part.
- Add `RunOptions.Clock quartz.Clock` with nil-safe default.
- Tag the startup guard timer as `"startupGuard"` for quartz trap
targeting.
- Rewrite the four startup-timeout integration tests to use
`quartz.NewMock(t)` with trap/advance/release sequences instead of
wall-clock sleeps.
- Add `awaitRunResult` helper so tests fail with a clear message instead
of hanging when `Run` does not complete.
Closes https://github.com/coder/internal/issues/1460
Adds `coder exp chat context add` and `coder exp chat context clear`
commands that run inside a workspace to manage chat context files via
the agent token.
`add` reads instruction and skill files from a directory (defaulting to
cwd) and inserts them as context-file messages into an active chat.
Multiple calls are additive — `instructionFromContextFiles` already
accumulates all context-file parts across messages.
`clear` soft-deletes all context-file messages, causing
`contextFileAgentID()` to return `!found` on the next turn, which
triggers `needsInstructionPersist=true` and re-fetches defaults from the
agent.
Both commands auto-detect the target chat via `CODER_CHAT_ID` (already
set by `agentproc` on chat-spawned processes), or fall back to
single-active-chat resolution for the agent. The `--chat` flag overrides
both.
Also adds sub-agent context inheritance: `createChildSubagentChat` now
copies parent context-file messages to child chats at spawn time, so
delegated sub-agents share the same instruction context without
independently re-fetching from the workspace agent.
<details><summary>Implementation details</summary>
**New files:**
- `cli/exp_chat.go` — CLI command tree under `coder exp chat context`
**Modified files:**
- `agent/agentcontextconfig/api.go` — `ConfigFromDir()` reads context
from an arbitrary directory without env vars
- `codersdk/agentsdk/agentsdk.go` — `AddChatContext`/`ClearChatContext`
SDK methods
- `coderd/workspaceagents.go` — POST/DELETE handlers on
`/workspaceagents/me/chat-context`
- `coderd/coderd.go` — Route registration
- `coderd/database/queries/chats.sql` — `GetActiveChatsByAgentID`,
`SoftDeleteContextFileMessages`
- `coderd/database/dbauthz/dbauthz.go` — RBAC implementations for new
queries
- `coderd/x/chatd/subagent.go` — `copyParentContextFiles` for sub-agent
inheritance
- `cli/root.go` — Register `chatCommand()` in `AGPLExperimental()`
**Auth pattern:** Uses `AgentAuth` (same as `coder external-auth`) —
agent token via `CODER_AGENT_TOKEN` + `CODER_AGENT_URL` env vars.
</details>
> 🤖 Generated by Coder Agents
---------
Co-authored-by: Michael Suchacz <203725896+ibetitsmike@users.noreply.github.com>
The agents chat interface displays thumbnails for videos recorded by the
computer use agent. Currently, to display a thumbnail, the frontend
downloads the entire video and shows the first frame. This PR starts
storing a new thumbnail file in the database for every recorded video,
and exposes the file id in the `wait_agent` tool result alongside the
recording file id, so the frontend can fetch just the thumbnail.
## Problem
Workspaces showed as "Healthy" immediately after creation while the
agent was still downloading, starting, or connecting. If the agent never
connected, the workspace stayed "Healthy" for the entire connection
timeout (~120s), then abruptly flipped to "Unhealthy".
## Root cause
In `db2sdk.WorkspaceAgent`, the health switch had no case for
`WorkspaceAgentConnecting`. Agents in `connecting` status with a
non-`off` lifecycle (e.g. `created` after a fresh build) fell through to
the `default` case and were marked `Healthy = true`.
## Fix
Add an explicit case for `WorkspaceAgentConnecting` that sets `Healthy =
false` with reason `"agent has not yet connected"`. The case is placed
after the existing `!connected + off` case (which correctly catches
stopped agents as "not running") and before the `timeout`/`disconnected`
cases.
```
Status + Lifecycle → Health reason
──────────────────────────────────────────────────────
any !connected + off → "agent is not running"
connecting + created/starting → "agent has not yet connected" ← NEW
timeout + any → "agent is taking too long to connect"
disconnected + any → "agent has lost connection"
connected + start_error → "agent startup script exited with an error"
connected + shutting_down → "agent is shutting down"
connected + ready/starting → healthy
```
The frontend already handles this case — `getAgentHealthIssue()` returns
"Workspace agent is still connecting" with `severity: "info"` for
unhealthy workspaces with connecting agents.
## Test changes
- **Healthy test**: now actually connects the agent via `agenttest.New`
before asserting health (previously passed due to the bug).
- **New Connecting test**: verifies a never-connected agent is correctly
marked unhealthy.
- **Mixed health test**: connects a1 and waits for the mixed state
(`a1.Healthy && !workspace.Healthy`) to avoid a race where both agents
are initially connecting.
- **Sub-agent excluded test**: connects the parent agent and waits for
it to be healthy before creating the sub-agent.
- **TestWorkspaceAgent/Connect**: flipped assertion to `Health.Healthy
== false` for a `dbfake` agent that never connects.
<details>
<summary>Review notes</summary>
### Known follow-up
The `healthy:false` workspace search filter maps to `[disconnected,
timeout]` and does not include `connecting`. This is a pre-existing gap
that is now more consequential — a workspace unhealthy solely due to a
connecting agent won't appear in `healthy:false` results. Worth a
follow-up issue.
### Deep review findings addressed
| Finding | Severity | Status |
|---------|----------|--------|
| Mixed health test race (all 3 reviewers) | P2 | Fixed — tightened
`Eventually` condition |
| `TestWorkspaceAgent/Connect` assertion break | P1 | Fixed — flipped
assertion |
| CLI renders red for connecting agents | Obs | Acknowledged — design
trade-off, accurate but visually strong for transient state |
| Switch case ordering overlap | Obs | Documented with inline comment |
</details>
> 🤖 This PR was created with the help of Coder Agents, and needs a human
review. 🧑💻
Workspace agent logs could still fail after the earlier invalid UTF-8
fix because NUL bytes are valid Go/protobuf strings but are rejected by
Postgres text columns. The legacy HTTP log upload path also bypassed the
old sanitization entirely, and both server insert paths computed
logs_length from the unsanitized input.
Add a shared log-output sanitizer in agentsdk, use it in the protobuf
conversion path and both server-side insert paths, and compute
OutputLength from the sanitized string so overflow accounting matches
what is actually stored. This keeps the old invalid UTF-8 behavior while
also handling embedded NUL bytes consistently across DRPC and HTTP log
ingestion.
Refs [#23292 ](https://github.com/coder/coder/issues/23292)
Refs [#13433 ](https://github.com/coder/coder/issues/13433)
## Summary
Adds `credential_kind` and `credential_hint` columns to
`aibridge_interceptions` to record how each LLM request was
authenticated and provide a masked credential identifier for audit
purposes.
This enables admins to distinguish between centralized API keys,
personal API keys, and subscription-based credentials in the
interceptions audit log.
## Changes
- New migration adding `credential_kind`and `credential_hint` to
`aibridge_interceptions`
- Updated `InsertAIBridgeInterception` query and proto definition to
carry the new fields
- Wired proto fields through `translator.go` and `aibridgedserver.go` to
the database
Depends on https://github.com/coder/aibridge/pull/239
Adds an optional `CreatedAt` timestamp to `tool-call` and `tool-result`
`ChatMessagePart` variants so the frontend can compute tool execution
duration (`result.created_at - call.created_at`).
Timestamps are recorded at the correct moments in the chatloop:
- **Tool-call**: when the model stream emits the tool call
- **Tool-result**: when tool execution completes (or is interrupted)
These are passed through `PersistedStep.PartCreatedAt` so the
persistence layer can apply accurate timestamps to stored parts.
SSE-published parts also carry `CreatedAt` for real-time display.
Old persisted messages without `created_at` deserialize to `nil` — fully
backward compatible.
<details><summary>Implementation notes (Coder Agents
generated)</summary>
### Why not stamp in `PartFromContent`?
`PartFromContent` is called both for SSE publishing (correct timing) and
during persistence (wrong timing — both tool-call and tool-result would
get the same "persistence time" timestamp, yielding ~0 duration).
Instead, timestamps are captured in the chatloop at the right moments
and carried through `PersistedStep.PartCreatedAt` as a
`map[string]time.Time` keyed by `"call:<id>"` / `"result:<id>"`.
### Interrupted tool calls
`persistInterruptedStep` also stamps `CreatedAt` on synthetic error
results for cancelled/interrupted tool calls, so partial duration is
available.
### Files changed
| File | Change |
|------|--------|
| `codersdk/chats.go` | Add `CreatedAt *time.Time` field |
| `codersdk/chats_test.go` | JSON round-trip test |
| `coderd/database/dbtime/dbtime.go` | Add `TimePtr` helper |
| `coderd/x/chatd/chatloop/chatloop.go` | Track timestamps, pass through
`PersistedStep` |
| `coderd/x/chatd/chatd.go` | Apply timestamps during persistence |
| `coderd/x/chatd/chatprompt/chatprompt_test.go` | Verify
`PartFromContent` does NOT stamp |
| `site/src/api/typesGenerated.ts` | Auto-generated |
</details>
---------
Co-authored-by: Ethan <39577870+ethanndickson@users.noreply.github.com>
Adds client-executed dynamic tools to the chat API. Dynamic tools are
declared by the client at chat creation time, presented to the LLM
alongside built-in tools, but executed by the client rather than chatd.
This enables external systems (Slack bots, IDE extensions, Discord bots,
CI/CD integrations) to plug custom tools into the LLM chat loop without
modifying chatd's built-in tool set.
Modeled after OpenAI's Assistants API: the chat pauses with
`requires_action` status when the LLM calls a dynamic tool, the client
POSTs results back via `POST /chats/{id}/tool-results`, and the chat
resumes.
See [this example](https://github.com/coder/coder-slackbot-poc) as a
reference for how this is used. It's highly-configurable, which would
enable creating chats from webhooks, periodically polling, or running as
a Slackbot.
<details>
<summary>Design context</summary>
### Architecture
The chatloop **exits** when it encounters dynamic tools and
**re-enters** when results arrive. No blocking channels, no pubsub for
tool results, no in-memory registry. The DB is the only coordination
mechanism.
```
Phase 1 (chatloop):
LLM response → execute built-in tools only →
Persist(assistant + built-in results) →
status = requires_action → chatloop exits
Phase 2 (POST /tool-results):
Persist(dynamic tool results) →
status = pending → wakeCh → chatloop re-enters
```
### Validation (POST /tool-results)
1. Chat status must be `requires_action` (409 if not)
2. Read chat's `dynamic_tools` → set of dynamic tool names
3. Read last assistant message → extract tool-call parts matching
dynamic tool names
4. Submitted tool_call_ids must match exactly (400 for missing/extra)
5. Persist tool-result message parts, set status to `pending`, signal
wake
### Idempotency
Tool call IDs scoped per LLM step. State machine (`requires_action` →
`pending`) is the guard. First POST wins, subsequent get 409.
### Mixed tool calls
When the LLM calls both built-in and dynamic tools in one step, built-in
tools execute immediately. Their results are persisted in phase 1.
Dynamic tool results arrive via POST in phase 2. The LLM sees all
results when the chatloop resumes.
</details>
> 🤖 Generated by Coder Agents
Adds telemetry collection for the agents chat system (`/agents`) to the
existing telemetry snapshot pipeline.
Three new snapshot fields:
- **`Chats`** — per-chat metadata (id, owner, status, mode,
workspace_id, root_chat_id, has_parent, archived, model config)
collected time-windowed via `createdAfter`
- **`ChatMessageSummaries`** — per-chat aggregated message metrics
(counts by role, token sums by type, cost, runtime, model count,
compression count) collected time-windowed
- **`ChatModelConfigs`** — model configuration metadata (provider,
model, context limit, enabled, default) collected as full dump
No PII is included — titles, message content, and URLs are excluded at
the SQL level. Only structural metadata flows through telemetry.
<details><summary>Implementation plan</summary>
### SQL Queries (`coderd/database/queries/chats.sql`)
- `GetChatsCreatedAfter` — time-windowed chat metadata
- `GetChatMessageSummariesPerChat` — per-chat message aggregates via
`GROUP BY`
- `GetChatModelConfigsForTelemetry` — full dump of model configs
### Telemetry (`coderd/telemetry/telemetry.go`)
- `Chat`, `ChatMessageSummary`, `ChatModelConfig` structs
- `ConvertChat`, `ConvertChatMessageSummary`, `ConvertChatModelConfig`
conversion functions
- Three `eg.Go()` blocks in `createSnapshot()` following the existing
collection pattern
### Authorization (`coderd/database/dbauthz/dbauthz.go`)
- System-only access for all three queries via `rbac.ResourceSystem`
### Tests
- `TestChatsTelemetry` in `coderd/telemetry/telemetry_test.go` — creates
chats (root + child), messages with token/cost data, model configs;
verifies all snapshot fields
- dbauthz test entries for all three queries in
`coderd/database/dbauthz/dbauthz_test.go`
</details>
> 🤖 Generated by Coder Agents
Fixes https://github.com/coder/coder/issues/23910
Adds periodic cleanup of chats and chat files to the dbpurge background
goroutine, with a configurable retention period exposed in the Agent
settings UI.
> 🤖 Written by a Coder Agent. Reviewed by a human.
Update queries as prep work for user secrets API development:
- Switch all lookups and mutations from ID-based to user_id + name
- Split list query into metadata-only (for API responses) and
with-values (for provisioner/agent)
- Add partial update support using CASE WHEN pattern for write-only
value fields
- Include value_key_id in create for dbcrypt encryption support
- Update dbauthz wrappers and remove stale methods from dbmetrics
## Summary
Replaces N per-chat heartbeat goroutines with a single centralized
heartbeat loop that issues one `UPDATE` per 30s interval for all running
chats on a worker.
## Problem
Each running chat spawned a dedicated goroutine that issued an
individual `UPDATE chats SET heartbeat_at = NOW() WHERE id = $1 AND
worker_id = $2 AND status = 'running'` query every 30 seconds. At 10,000
concurrent chats this produces **~333 DB queries/second** just for
heartbeats, plus ~333 `ActivityBumpWorkspace` CTE queries/second from
`trackWorkspaceUsage`.
## Solution
New `UpdateChatHeartbeats` (plural) SQL query replaces the old singular
`UpdateChatHeartbeat`:
```sql
UPDATE chats
SET heartbeat_at = @now::timestamptz
WHERE worker_id = @worker_id::uuid
AND status = 'running'::chat_status
RETURNING id;
```
A single `heartbeatLoop` goroutine on the `Server`:
1. Ticks every `chatHeartbeatInterval` (30s)
2. Issues one batch UPDATE for all registered chats
3. Detects stolen/completed chats via set-difference (equivalent of old
`rows == 0`)
4. Calls `trackWorkspaceUsage` for surviving chats
`processChat` registers an entry in the heartbeat registry instead of
spawning a goroutine.
## Impact
| Metric | Before (10K chats) | After (10K chats) |
|---|---|---|
| Heartbeat queries/sec | ~333 | ~0.03 (1 per 30s per replica) |
| Heartbeat goroutines | 10,000 | 1 |
| Self-interrupt detection | Per-chat `rows==0` | Batch set-difference |
---
> 🤖 Generated by Coder Agents
<details><summary>Implementation notes</summary>
- Uses `@now` parameter instead of `NOW()` so tests with `quartz.Mock`
can control timestamps.
- `heartbeatEntry` stores `context.CancelCauseFunc` + workspace state
for the centralized loop.
- `recoverStaleChats` is unaffected — it reads `heartbeat_at` which is
still updated.
- The old singular `UpdateChatHeartbeat` is removed entirely.
- `dbauthz` wrapper uses system-level `rbac.ResourceChat` authorization
(same pattern as `AcquireChats`).
</details>
Audit and connection log pages were timing out due to expensive COUNT(*)
queries over large tables. This commit adds opt-in count capping: requests can
return a `count_cap` field signaling that the count was truncated at a threshold,
avoiding full table scans that caused page timeouts.
Text-cast UUID comparisons in regosql-generated authorization queries
also contributed to the slowdown by preventing index usage for connection
and audit log queries. These now emit native UUID operators.
Frontend changes handle the capped state in usePaginatedQuery and
PaginationWidget, optionally displaying a capped count in the pagination
UI (e.g. "Showing 2,076 to 2,100 of 2,000+ logs")
Related to:
https://linear.app/codercom/issue/PLAT-31/connectionaudit-log-performance-issue
> This PR was authored by Mux on behalf of Mike.
External MCP tools returned by `ConnectAll` were ordered by goroutine
completion, making the tool list nondeterministic across chat turns.
This broke prompt-cache stability since tools are serialized in order.
Sort tools by their model-visible name after all connections complete,
matching the existing pattern in workspace MCP tools
(`agent/x/agentmcp/manager.go`). Also guards against a nil-client panic
in cleanup when a connected server contributes zero tools after
filtering.
Needed by #23833
Adds a `chat_file_links` association table to track which files are
associated with each chat.
- `AppendChatFileIDs` query links a file to a chat with deduplication
- `GetChatFileMetadataByIDs` query returns lightweight file metadata by
IDs
- Tool-created files (e.g. `propose_plan`) are linked to the chat after
insert
- User-uploaded files are linked to the chat when the referencing
message is sent
- Single-chat GET endpoint hydrates `files: ChatFileMetadata[]` on the
response
> 🤖 Created by Coder Agents and massaged into shape by a human.
Kyle Carberry